pubs.acs.
org/joc
A Concise and Convergent Synthesis of PA-824
Maurice A. Marsini, Paul J. Reider,* and Erik J. Sorensen*
Department of Chemistry, Frick Chemical Laboratory,
Princeton University, Princeton, New Jersey 08544,
United States
preider@princeton.edu; ejs@princeton.edu
Received August 11, 2010
An efficient four-step synthesis of PA-824, a promising
antituberculosis drug candidate, has been developed. This
concise approach offers significant improvements over the
synthetic route currently used for large-scale production.
Tuberculosis (TB) is a devastating bacterial infection that
kills more than 1.8 million people each year worldwide.1 Over
98% of TB fatalities result from one specific bacterial strain,
Mycobacterium tuberculosis (Mtb), where the presence of a
single bacterium can propagate a fatal infection. Although
Mtb was identified nearly 130 years ago and antibiotics have
been available to treat TB since the 1940s, this pathogen remains
a persistent problem because of the emergence of multidrug
resistant (MDR-TB) and extremely drug resistant (XDR-TB)
strains, as well as the lack of resources available to provide
drugs to endemic areas. As the FDA has not approved a new
TB-specific drug in over 40 years, there exists a critical need for
the development of new TB treatments that can be deployed on
a large scale to developing areas of the world where the disease
is most harmful.2
Currently, there are at least ten anti-TB drug candidates
with varying biological mechanisms of action being evalu-
ated in clinical trials.3 These therapies fit into two general
classes: (1) those already used in first- and second-line
treatment of active TB, and (2) those with novel mechanisms
of action against both replicating and nonreplicating TB.
Among the second group, bicyclic nitroimidazoles have
emerged as a promising structural class (Figure 1).4 While
(1) (a) World Health Organization Fact sheet 104, March 2010. (b) Dye,
C.; Scheele, S.; Dolin, P.; Pathania, V.; Raviglione, M. C. J. Am. Med. Assoc.
1999, 282, 677–686.
(2) Duncan, K. Tuberculosis 2003, 83, 201–207.
(3) Ginsberg, A. M. Tuberculosis 2010, 90, 162–167.
(4) Barry, C. E.; Boshoff, H. I. M.; Dowd, C. S. Curr. Pharm. Des. 2004,
10, 3239–3262.
(5) Stover, C. K.; Warrener, P.; VanDevanter, D. R.; Sherman, D. R.;
Arain, T. M.; Langhorne, M. H.; Anderson, S. W.; Towell, J. A.; Yuan, Y.;
McMurray, D. N.; Kreiswirth, B. N.; Barry, C. E.; Baker, W. R. Nature 2000,
405, 962–966.
DOI: 10.1021/jo1015807 Published on Web 10/07/2010
r 2010 American Chemical Society
metronidazole (2) is widely used as a general antibacterial
treatment, PA-824 (1) stands out as an exciting experimental
anti-TB drug candidate.5 PA-824 was found to inhibit TB
with minimum inhibitory concentration (MIC) values of
0.015-0.25 μg/mL.6 Furthermore, several strains of MDR-
TB exhibited comparable susceptibility to PA-824, indicat-
ing a lack of potential cross-resistance with current MDR-
TB therapies.7 PA-824 is being developed by the Global
Alliance for TB Drug Development (GATB) and is now in
advanced phase II clinical trials.8
FIGURE 1. Nitroimidazole drug candidates for tuberculosis.
In replicating TB, compounds 1 and 3 are believed to
inhibit Mtb cell wall formation by disrupting mycolic acid
synthesis.9 However, the primary mechanism of action in
both replicating and nonreplicating TB is believed to involve
reduction of the imidazole ring by Fgd1, an F420-dependent
glucose-6-phosphate dehydrogenase, and Rv3547, a deaza-
flavin-dependent nitroreductase, after passage through the
bacterial cell membrane to afford the des-nitro compound
and nitric oxide (NO).10 Cyclic voltammetry studies11 have
since justified and guided the synthesis and biological in-
vestigation of several analogues of PA-824 (1) and OPC-
67683 (3).12
(6) Lenaerts, A. J.; Gruppo, V.; Marietta, K. S.; Johnson, C. M.; Driscoll,
D. K.; Tompkins, N. M.; Rose, J. D.; Reynolds, R. C.; Orme, I. M.
Antimicrob. Agents Chemother. 2005, 49, 2294–2301.
(7) (a) Nuermberger, E.; Rosenthal, I.; Tyagi, S.; Williams, K. N.;
Almeida, D.; Peloquin, C. A.; Bishai, W. R.; Grosset, J. H. Antimicrob.
Agents Chemother. 2006, 50, 2621–2625. (b) Tyagi, S.; Nuermberger, E.;
Yoshimatsu, T.; Williams, K.; Rosenthal, I.; Lounis, N.; Bishai, W.; Grosset,
J. Antimicrob. Agents Chemother. 2005, 49, 2289–2293. (c) Hu, Y.; Coates,
A. R. M.; Mitchison, D. A. Int. J. Tuberc. Lung Dis. 2008, 12, 69–73.
(8) Global Alliance for TB Drug Development report, October 26, 2008
(http://www.tballiance.org/newscenter).
(9) Yuan, Y.; Barry, C. E. Proc. Natl Acad. Sci. U.S.A. 1996, 93, 12828–
12833.
(10) (a) Manjunatha, U.; Boshoff, H. I. M.; Barry, C. E. Commun. Integr.
Biol. 2009, 2, 215–218. (b) Maroz, A.; Shinde, S. S.; Franzblau, S. G.; Ma, Z.;
Denny, W. A.; Palmer, B. D.; Anderson, R. F. Org. Biomol. Chem. 2010, 8,
413–418. (c) Manjunatha, U. H.; Boshoff, H. I. M.; Dowd, C. S.; Zhang, L.;
Albert, T. J.; Norton, J. E.; Daniels, L.; Dick, T.; Pang, S. S.; Barry, C. E.
Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 431–436. (d) Singh, R.; Manjunatha,
U.; Boshoff, H. I. M.; Ha, Y. H.; Niyomrattanakit, P.; Ledwidge, R.; Dowd,
C. S.; Lee, I. Y.; Kim, P.; Zhang, L.; Kang, S.; Keller, T. H.; Jiricek, J.; Barry,
C. E. Science 2008, 322, 1392–1395. (e) Anderson, R. F.; Shinde, S. S.;
Maroz, A.; Boyd, M.; Palmer, B. D.; Denny, W. A. Org. Biomol. Chem. 2008,
6, 1973–1980. (f) Bashiri, G.; Squire, C. J.; Moreland, N. J.; Baker, E. N.
J. Biol. Chem. 2008, 283, 17531–17541.
(11) (a) Yanez, C.; Bollo, S.; Nunez-Vergara, L. J.; Squella, J. A. Anal.
Lett. 2001, 34, 2335–2348. (b) Wardman, P.; Clarke, E. D. J. Chem. Soc.,
Faraday Trans. 1 1976, 72, 1377–1390. (c) Bollo, S.; Nunez-Vergara, L. J.;
Squella, J. A. J. Electroanal. Chem. 2004, 562, 9–14.
J. Org. Chem. 2010, 75, 7479–7482 7479
JOC Note
SCHEME 1. Original Production Process for PA-824a
a
DHP = 3,4-dihydropyran; p-TsOH = p-toluenesulfonic acid;
MsOH=methanesulfonic acid.
The original synthesis used to obtain material for clinical
trials provides PA-824 in five linear steps from 2,4-dinitroi-
midazole 4 (Scheme 1).13 Unfortunately, dinitroimidazole 4
is explosive, and both 4 and the starting epoxide 5 are
expensive for use on large scale. Furthermore, the synthesis
requires four chromatographic separations and an inefficient
protection-deprotection sequence in order to selectively
construct the oxazine core that are not suitable for use in a
kilogram-scale process. A solventless modification of this
pathway has also been developed.14
Since the overall cost of any anti-TB drug is a major factor
in future production and necessitates a safe, efficient, and
economic synthesis of this active compound, we set out to
develop a synthesis of PA-824 that would address these
concerns and be amenable to large-scale process develop-
ment (Scheme 2).
Our key tactical consideration hinges upon the direct,
convergent coupling of a suitably safe alternative to
2,4-dinitroimidazole (A) and an appropriately functionalized
pseudosymmetrical glycidol derivative (B). Ideally, glycidol
B would contain the essential p-(trifluoromethoxy)benzyl
ether moiety with the correct stereochemical configuration
as well as functional groups to enable either a direct or stepwise
construction of the central six-membered oxazine ring.
Our studies began with the synthesis of a suitable glycidol
derivative (Scheme 3). (R)-3-Chloro-1,2-propanediol (9),
which is commercially available and easily obtainable from the
hydrolytic kinetic resolution (HKR)15 of racemic epichlorohy-
drin on ton-scale, is selectively benzoylated to afford chlor-
ohydrin 10 in 78-82% yield and >99% ee without the need for
further purification. We chose the p-methoxybenzoyl group as
(12) (a) Li, X.; Manjunatha, U. H.; Goodwin, M. B.; Knox, J. E.;
Lipinski, C. A.; Keller, T. H.; Barry, C. E.; Dowd, C. S. Bioorg. Med. Chem.
Lett. 2008, 18, 2256–2262. (b) Nagarajan, K.; Shankar, R. G.; Rajappa, S.;
Shenoy, S. J.; Costa-Pereira, R. Eur. J. Med. Chem. 1989, 24, 631–633. (c)
Matsumoto, M.; Hashizume, H.; Tomishige, T.; Kawasaki, M.; Tsubouchi,
H.; Sasaki, H.; Shimokawa, Y.; Komatsu, M. PLoS Med. 2006, 3, 2131–
2144. (d) Sutherland, H. S.; Blaser, A.; Kmentova, I.; Franzblau, S. G.; Wan,
B.; Wang, Y.; Ma, Z.; Palmer, B. D.; Denny, W. A.; Thompson, A. M.
J. Med. Chem. 2010, 53, 855-866 and references cited therein.
(13) Baker, W. R.; Shaopei, C.; Keeler, E. L. Nitro-[2,1-b]imidazopyran
Compounds and Antibacterial Uses Thereof. U.S. Patent 6,087,358, 2000.
(14) Orita, A.; Miwa, K.; Uehara, G.; Otera, J. Adv. Synth. Catal. 2007,
349, 2136–2144.
(15) (a) Tokunaga, M.; Larrow, J. F.; Kakiuchi, F.; Jacobsen, E. N.
Science 1997, 277, 936–938. (b) Furrow, M. E.; Schaus, S. E.; Jacobsen, E. N.
J. Org. Chem. 1998, 63, 6776–6777.
(16) Corey, E. J.; Guzman-Perez, A.; Noe, M. C. J. Am. Chem. Soc. 1995,
117, 10805–10816.
7480 J. Org. Chem. Vol. 75, No. 21, 2010
Marsini et al.
SCHEME 2. General Synthetic Strategy
SCHEME 3. Synthesis of a Functionalized Glycidol Derivativea
a
Cl3CCN = trichloroacetonitrile; TBME = tert-butylmethyl ether;
TfOH = trifluoromethanesulfonic acid.
a removable trigger for our proposed oxazine construction
because of its stability and migratory resistance.16 To circum-
vent epoxide formation, benzylation of 10 was accomplished
under acidic conditions utilizing the benzyltrichloroacetimi-
date derived from p-(trifluoromethoxy)benzyl alcohol 11 and
stoichiometric trifluoromethanesulfonic acid in CH2Cl2 at
0 °C in 80% yield. This sequence assembles the desired chloride
12 in two linear steps from chloride 9 and 60% overall yield.
While the glycidol-derived alkyl chloride 12 was antici-
pated to be relatively unreactive as an alkylating agent, we
observed that the desired N-alkylation of 2-chloro-4-nitroi-
midazole 13 (CNI)17 could be achieved by the action of
sodium iodide and potassium carbonate on a mixture of 12
and 13 in DMF at 120 °C (Scheme 4). This union was com-
pletely regioselective, most likely due to the bulkiness of the
alkylating agent and higher reaction temperature, and af-
forded the desired product 14 in yields ranging from 40%
to 50% after recrystallization of the crude reaction mixture.
This alkylation step offers a significant advantage to the
existing route by circumventing the need for an excess of
expensive reagents to obtain the desired regioselectivity.
(17) (a) 2-Chloro-4-nitroimidazole 13 (CNI) is a non-explosive, crystal-
line compound that is commercially available from Asahi Kasei. For
information concerning the preparation of 2-halo-4-nitroimidazoles and
intermediates thereof, see: Wuellner, G.; Herkenrath, F.-W.; Juelich, A.;
Yamada, Y.; Kawabe, S. PCT Int. Appl. WO 2010021409, August 21, 2009.
(b) For an example of the use of this reagent in the synthesis of PA-824
analogues, please see: Kim, P.; Zhang, L.; Manjunatha, R. S.; Patel, S.;
Jiricek, J.; Barry, C. E.; Dowd, C. S. J. Med. Chem. 2009, 12, 1317–1328.
Marsini et al.
SCHEME 4. Synthesis of PA-824
The key annulation of the oxazine ring could then be
achieved by exposing compound 14 to 3 equiv of potassium
hydroxide in methanol at 0 °C; this simple procedure effected
saponification of the benzoate ester and, after warming to room
temperature, anionic cyclization-elimination to produce PA-824
(1) in 60-70% yield and >99.9% ee after recrystallization.
The hallmarks of this efficient, concise synthesis of PA-824
include the convergent design, simplicity of reagents and
experimental procedures, the utilization of the safe imidazole
derivative 13, and easy accessible chiral building block 12.
The suitability of this process for the kilogram-scale produc-
tion of Pa-824 is being investigated.
Experimental Section
(R)-3-Chloro-2-hydroxypropyl-4-methoxybenzoate 10. To a
stirred solution of (R)-3-chloro-1,2-propanediol (6.61 g, 59.8
mmol) in CH2Cl2 (120 mL) was added imidazole (4.07 g, 59.8
mmol). After the reaction mixture had cooled to 0 °C, p-metho-
xybenzoyl chloride (10.2 g, 59.8 mmol) in CH2Cl2 (10 mL) was
added dropwise via addition funnel. The resulting solution
was allowed to warm to room temperature and stirred until
complete consumption of starting material by thin layer chro-
matography (TLC). The mixture was poured into saturated aq
NH4Cl (150 mL), and the aqueous layer was extracted with
CH2Cl2 (3
100 mL). The combined organic extracts were dried
(MgSO4), filtered, and concentrated in vacuo to provide chloro-
hydrin 10 as a clear, viscous oil (11.43 g, 78%) that was used
without further purification. Rf 0.3 (20% EtOAc/hexanes);
ee >99% as determined by chiral SFC (see the Supporting
Information); 1H NMR (500 MHz, CDCl3) δ 8.07-7.94 (m,
2H), 7.00-6.88 (m, 2H), 4.46 (d, J = 5.1 Hz, 2H), 4.22 (dd, J =
10.6, 5.3 Hz, 1H), 3.88 (s, 3H), 3.78-3.64 (m, 2H), 2.73 (d, J =
5.6 Hz, 1H); 13C NMR (126 MHz, CDCl3) δ 166.7, 163.9, 132.1,
121.9, 114.0, 70.1, 65.7, 55.7, 46.3; IR [CH2Cl2 solution] νmax
(cm-1) 3454, 2959, 2889, 1713, 1606, 1512, 1259, 1170, 1104,
1028, 848, 770, 697, 614; HRMS (ESI-TOF) calcd for
C11H13ClO4 (M)þ 244.0573, found 244.0501.
4-(Trifluoromethoxy)benzyl Trichloroacetimidate. To a stir-
red suspension of NaH (60%, 127 mg, 3.45 mmol) in t-BuOMe
(100 mL) was added 4-(trifluoromethoxy)benzyl alcohol 11
(5 mL, 34.5 mmol) dropwise by syringe at room temperature.
After H2 evolution had ceased (about 5 min), the clear mixture
was cooled to 0 °C. Cl3CCN (3.46 mL, 34.5 mmol) was added
dropwise by syringe, and the resulting solution was stirred for
1 h. After warming to room temperature, the solution was con-
centrated in vacuo, and the residue was resuspended in heptane
(100 mL) containing MeOH (0.14 mL, 3.45 mmol). After 10 min
JOC Note
of stirring, the suspension was filtered through a thin bed of
silica and washed with heptane. Concentration in vacuo affords
the title compound as a yellow-orange liquid (11.4 g, 98%) that
was used directly without further purification. Rf 0.3 (5%
EtOAc/hexanes); 1H NMR (501 MHz, CDCl3) δ 8.43 (s, 1H),
7.48 (d, J = 8.2 Hz, 2H), 7.24 (d, J = 8.2 Hz, 2H), 5.35 (s, 2H);
13
C NMR (126 MHz, CDCl3) δ 162.6, 149.3, 134.3, 129.4, 121.6,
121.3, 119.6, 69.9; IR [CH2Cl2 solution] νmax (cm-1) 3345, 2954,
1667, 1511, 1261, 1221, 1166, 1077, 998, 828, 797, 649; HRMS
(ESI-TOF) calcd for C10H7Cl3F3NO2 334.95, found C8H6F3O
174.03.
(R)-3-Chloro-2-(4-(trifluoromethoxy)benzyloxy)propyl 4-Meth-
oxybenzoate 12. To a stirred solution of chlorohydrin 10 (7.15 g,
29 mmol) and 4-(trifluoromethoxy)benzyl trichloroacetimi-
date (9.76 g, 29 mmol) in CH2Cl2 (58 mL) at 0 °C was added
TfOH (1.95 mL, 29 mmol) dropwise via glass syringe. The re-
sulting mixture was allowed to slowly warm to room tempera-
ture and stirred overnight. After completion, the mixture was
quenched slowly with saturated aq NaHCO3 until gas evolu-
tion ceased. The layers were separated, and the aqueous layer
was extracted once with CH2Cl2 (100 mL). The combined
organic extracts were dried (MgSO4), filtered, and concen-
trated in vacuo. Chromatography (5% EtOAc/hexanes) af-
fords chloride 12 (9.7 g, 80%) as a clear, light yellow oil. Rf 0.2
(5% EtOAc/hexanes); 1H NMR (501 MHz, CDCl3) δ 8.00 (d,
J = 8.9 Hz, 2H), 7.42 (d, J = 8.5 Hz, 2H), 7.19 (d, J = 8.2 Hz,
2H), 6.95 (d, J = 8.9 Hz, 2H), 4.72 (d, J = 13.6 Hz, 2H), 4.63-
4.40 (m, 2H), 3.99 (t, J = 5.1 Hz, 1H), 3.90 (s, 3H), 3.79-3.65
(m, 2H); 13C NMR (126 MHz, CDCl3) δ 166.1, 163.8, 136.5,
131.9, 129.4, 122.1, 121.7, 121.2, 113.9, 110.2, 77.0, 71.7, 63.5,
55.7, 43.6; IR [CH2Cl2 solution] νmax (cm-1) 2959, 2872, 2843,
1904, 1715, 1607, 1511, 1450, 1258, 1030, 921, 823, 770;
HRMS (ESI-TOF) calcd for C19H18ClF3O5 418.0795, found
418.0791.
(S)-3-(2-Chloro-4-nitro-1H-imidazol-1-yl)-2-(4-(trifluoro-meth-
oxy)benzyloxy)propyl 4-Methoxybenzoate 14. To a stirred solu-
tion of chloride 12 (17.4 g, 41.4 mmol) in DMF (42 mL) was
added 2-chloro-4-nitroimidazole (CNI, 13, from Asahi Kasei,
6.11 g, 41.4 mmol), K2CO3 (7.53 g, 53.9 mmol), and NaI (7.80 g,
53.9 mmol). The resulting suspension was then heated to 120 °C
(oil bath) and stirred at this temperature for 16 h. After comple-
tion, the reaction mixture was allowed to cool to room tempera-
ture and poured into ice/H2O. The mixture was extracted with
EtOAc (3
300 mL), and the combined organic extracts were
washed with brine, dried (Na2SO4), filtered, and concentrated in
vacuo. The resulting residue was directly recrystallized from
i-PrOH/hexanes to afford compound 14 (8.6 g, 41%) as a white
solid. Mp 109-112 °C; Rf 0.2 (1:2 EtOAc/hexanes); 1H NMR
(501 MHz, CDCl3) δ 7.98 (d, J = 8.9 Hz, 2H), 7.82 (s, 1H), 7.19
(dd, J = 24.3, 8.2 Hz, 2H), 6.96 (d, J = 8.8 Hz, 2H), 4.73 (d, J =
11.8 Hz, 2H), 4.47 (d, J = 4.8 Hz, 2H), 4.32 (d, J = 14.5 Hz, 2H),
4.17 (dd, J = 14.6, 8.5 Hz, 2H), 3.99 (s, 2H), 3.89 (s, 3H); 13C
NMR (126 MHz, CDCl3) δ 165.9, 164.2, 147.9, 135.2, 132.7,
132.0, 130.2, 129.7, 126.9, 122.0, 121.4, 114.2, 110.3, 74.8, 71.9,
62.1, 55.8, 49.7; IR [CH2Cl2 solution] νmax (cm-1) 2955, 1715,
1606, 1547, 1510, 1472, 1388, 134, 1258, 1221, 1168, 1029, 848;
HRMS (ESI-TOF) calcd for C22H19ClF3N3O7 529.0864, found
529.0859.
PA-824 (1). To a stirred suspension of compound 14 (6.0 g,
11.32 mmol) in anhydrous MeOH (18 mL) at 0 °C was added
ground KOH (1.91 g, 34 mmol). The mixture was pulled from
the bath and stirred at room temperature for 1 h. The resulting
suspension was poured into H2O and extracted with EtOAc (4

100 mL). The combined organic extracts were washed with brine,
dried (Na2SO4), filtered, and concentrated. Chromatography
(75% EtOAc/hexanes) followed by recrystallization (i-PrOH/
hexanes) affords PA-824 (1) (2.41 g, 62%) as a crystalline solid.
Mp 150-151 °C (lit.11a mp 149-150); Rf 0.2 (75% EtOAc/
J. Org. Chem. Vol. 75, No. 21, 2010 7481
JOC Note
hexanes); ee >99.9% as determined by chiral SFC (see the
Supporting Information); 1H NMR (500 MHz, d6-DMSO) δ
8.09 (s, 1H), 7.48 (d, J = 8.6 Hz, 2H), 7.39 (d, J = 8.2 Hz, 2H),
4.81-4.62 (m, 3H), 4.51 (d, J = 11.9 Hz, 1H), 4.39-4.19 (m, 3H);
13
C NMR (126 MHz, d6-DMSO) δ 148.7, 148.1, 143.0, 138.3,
130.4, 122.0, 120.0, 119.8, 69.7, 68.8, 67.51, 47.73; IR [CH2Cl2
solution] νmax (cm-1) 2877, 1580, 1543, 1509, 1475, 1404, 1380,
1342, 1281, 1221, 1162, 1116, 1053, 991, 904, 831, 740; HRMS
(ESI-TOF) calcd for C14H12F3N3O5 359.0729, found 359.0728.
Acknowledgment. We gratefully acknowledge Dr.
Christopher Cooper, Dr. Takushi Kaneko, and Dr. Zhenkun
7482 J. Org. Chem. Vol. 75, No. 21, 2010
Marsini et al.
Ma from the Global Alliance for TB Drug Development for
their gracious financial support of this research and valuable
discussions. We also thank the David Ju Foundation and
Amgen for generous financial support and Dr. Hiromu
Kawakubo (Asahi Kasei Chemicals Corporation) for a very
kind donation of CNI (13) used in these studies.
Supporting Information Available: 1H NMR and 13C NMR
spectra of 1 and 9-14 and chiral SFC data for 10 and 1. This
material is available free of charge via the Internet at http://
pubs.acs.org.